What's wrong with having multiple shields stacked up?
I see the display shield as independent from the others. RTC + eeprom + other I2C stuff can go on the second shield. And then, pressure, compass, temp and other gizmos on yet another shield.

One thing I don't understand is why the temperature sensor should be on the board. Doesn't the reading get affected by the heat dissipated by the surrounding components, especially when in an enclosure?

If you are considering DS1307 as your RTC, this is how it should look like:

Code:

#include <Wire.h>
#include <DS1307.h>


//--------------------------------------------------------------------------------------
/* DS1307 Real Time Clock - uses the wire library (I2C);
 *
 * RTC is set up through a separate sketch (+windows app talking on serial port).
 * I2C bus address of DS1307 is B1101000 (already defined in RTC.h).
 *
 * SETUP:     _ _                                
 *        X1-|oU |-Vcc
 *        X2-|   |-SQW/OUT square wave 1Hz output signal;
 *  Vbat(3V)-|   |-SCL     to Arduino analog 5
 *       GND-|   |-SDA     to Arduino analog 4
 *            ---
 *
 * Pull-up resistors (10k) are required on SDA (pins 5), SCL (pin 6), and SQW (pin 7).
 *
 *
 * References:
 *    http://www.datasheetcatalog.org/datasheet/maxim/DS1307.pdf
*/
//--------------------------------------------------------------------------------------

int tick = 0;


// interrupt handler for the 1Hz square wave generated by RTC (wired to arduino digital pin 3);
void SQW_1Hz_ISR()
{
  tick++;
  Serial.print("in ISR: ");
  Serial.println(tick);
}

void setTime()
{
  RTC.stop();
  RTC.set(DS1307_SEC, 20);
  RTC.set(DS1307_MIN, 07);
  RTC.set(DS1307_HR,  12);
  RTC.set(DS1307_DOW,  3);
  RTC.set(DS1307_DATE,28);
  RTC.set(DS1307_MTH,  1);
  RTC.set(DS1307_YR,   9);
}

void setup()
{
// initialize stuff here...

  // call this only once; comment out after time was set in RTC;
  setTime();

  Serial.begin(9600);

  // code courtesy of Mr. BroHogan (a.k.a. "Life Clock");
  Wire.beginTransmission(0x68);
  Wire.send(0x07);
  Wire.send(0x90);                       // 0x90=1Hz, 0x91=4kHz, 0x92=8kHz, 0x93=32kHz
  Wire.endTransmission();


  // define ISR for pin 3 (1Hz square wave from RTC);
  attachInterrupt(1, SQW_1Hz_ISR, FALLING);
}

Is this what you were looking for?

You can get some inspiration from the playground ("interfacing with hardware" section, look for DS1307, there are 2 entries).

The "DS1307" library I used was published on and downloaded from the arduino playground.

You can use the square wave output of the RTC to generate interrupts every second.

mac mini does not have an integrated monitor and its price is pretty close to the price of a netbook:

http://store.apple.com/us/browse/home/shop_mac/family/mac_mini?mco=MTE3MDI

It may also cost you more (in parts and labour)...

the LCD-Display needs at least 4 pins, each LED 2, the Speaker 2, each Antenna

Each LED would require just one output pin. 5 LEDs can be driven by only 3 pins.
Speaker would require, again, just one output pin, not 2. (If you are thinking the speaker has 2 connectors, the second connector would go to ground; same for the LEDs).

So then this seems to be a serial communication issue, not a device driver issue.
If I understand correctly, you can see the COM ports in the Arduino IDE, you can select them accordingly, but when you do the upload of the sketch, you get errors.
I am not familiar with iduino, but it may be that the bootloader uses a different baud rate than other standard bootloaders (those used in the original arduinos). Is this the only board you have? Can you try an arduino/freeduino/seeduino?

I don't think there is any confusion when creating virtual COM ports by the FTDI driver. The driver simply identifies the individual FTDI chip and creates a virtual COM port for it.
In Windows, you can see the virtual (USB) COM ports listed in Device Manager (System Properties/Hardware/Device Manager).
If you have 2 identical arduino boards (or for that matter, 2 identical FTDI chips), open Device Manager and expand the "Ports" node; then connect your first arduino/FTDI chip, and you will see a new port listed. Now disconnect the first and connect your second arduino/FTDI chip. You will see a new COM port listed and the old one gone. Smart way for the FTDI driver to manage multiple FTDI chips.

Sparkfun should be credited for creating at least 70 jobs. (Those people must be paid salaries.)

In case you were wondering what the result was:
http://timewitharduino.blogspot.com/2010/11/bookclock-contest-winner.html

Mowcius,
I did not expect you would not participate

Jeckson,
The idea is to be able to manufacture the box easily, cheaply and on a large scale, so it can be provided within a kit, eventually. (BroHogan put a lot of skill and effort into that box.)

I don't know how interesting (or not) this sounds, but I thought I would give it a try anyway: a contest for the "best" open-source plans for an enclosure. I've been struggling, with limited results, to design a case for one of my projects, the BookClock. This project uses the ubiquitous 8x32 LED display from Sure Electronics, so the appeal and benefits should be quite large. Ideally, this enclosure should be a book-sized wooden box, with the laser-cut sides, easy to assemble with minimal effort. But any other idea is welcome.

The "rules" and details of the contest can be found here:
http://timewitharduino.blogspot.com/2010/10/contest-enclosure-for-bookclock.html

The prize is a DWex watch.

was hoping not to go down the smd lane

That's almost impossible if you want to build a wrist watch, unless you have an unusually large wrist

The sonic ranger seems to have some relation with parallax ping sonic ranger

From the 2 photos posted, they look pretty different though.

I actually am looking for a photoresistor

Like this one?
http://www.adafruit.com/index.php?main_page=product_info&cPath=35&products_id=161